From: "Ilpo_Järvinen" <redacted>
Date: Fri, 1 Jun 2007 15:17:00 +0300 (EEST)

On Thu, 31 May 2007, David Miller wrote:

quoted

Is it possible to update fastpath_cnt_hint properly perhaps?

I think that would be valid and even accurate as it can checks skb's
seq against fastpath_skb_hint->seq. I'm in a hurry and will be a week
out of reach of internet connectivity but here's quick attempt to deal
with this cleanly. Compile tested (be extra careful with this one, it's 
done i a hurry :-)), consider to net-2.6. Considering the marginality of 
this issue, stable might really be an overkill for this one, only a
remotely valid concern comes to my mind in this case: possibility
of fackets_out > packet_out might not be dealt that cleanly everywhere 
(but I'm sure that you can come to a good decicion about it):

[PATCH] [TCP]: SACK fastpath did override adjusted fackets_out

Do same adjustment to SACK fastpath counters provided that
they're valid.

Signed-off-by: Ilpo Järvinen <redacted>

Since the case is marginal and the patch does need some testing
I'm putting this into net-2.6.23 for now.